TY - GEN
T1 - Belief-Space Planning Using Learned Models with Application to Underactuated Hands
AU - Kimmel, Andrew
AU - Sintov, Avishai
AU - Tan, Juntao
AU - Wen, Bowen
AU - Boularias, Abdeslam
AU - Bekris, Kostas E.
N1 - Funding Information:
The authors work was supported by NSF Awards 1734492 and 1723869. A. Kimmel and A. Sintov—Equal contribution.
Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022/1/1
Y1 - 2022/1/1
N2 - Acquiring a precise model is a challenging task for many important robotic tasks and systems - including in-hand manipulation using underactuated, adaptive hands. Learning stochastic, data-driven models is a promising alternative as they provide not only a way to propagate forward the system dynamics, but also express the uncertainty present in the collected data. Therefore, such models enable planning in the space of state distributions, i.e., in the belief space. This paper proposes a planning framework for solving Non-Observable Markov Decision Process (NOMDP) problems which employs learned stochastic models, expressing a distribution of states as a set of particles. The integration achieves anytime behavior in terms of returning paths of increasing quality under constraints for the probability of success to achieve a goal. The focus of this effort is on pushing the efficiency of the overall methodology despite the notorious computational hardness of belief-space planning. Experiments show that the proposed framework enables reaching a desired goal with higher success rate compared to alternatives in simple benchmarks. This work also provides an application to the motivating domain of in-hand manipulation with underactuated, adaptive hands, both in the case of physically-simulated experiments as well as demonstrations with a real hand.
AB - Acquiring a precise model is a challenging task for many important robotic tasks and systems - including in-hand manipulation using underactuated, adaptive hands. Learning stochastic, data-driven models is a promising alternative as they provide not only a way to propagate forward the system dynamics, but also express the uncertainty present in the collected data. Therefore, such models enable planning in the space of state distributions, i.e., in the belief space. This paper proposes a planning framework for solving Non-Observable Markov Decision Process (NOMDP) problems which employs learned stochastic models, expressing a distribution of states as a set of particles. The integration achieves anytime behavior in terms of returning paths of increasing quality under constraints for the probability of success to achieve a goal. The focus of this effort is on pushing the efficiency of the overall methodology despite the notorious computational hardness of belief-space planning. Experiments show that the proposed framework enables reaching a desired goal with higher success rate compared to alternatives in simple benchmarks. This work also provides an application to the motivating domain of in-hand manipulation with underactuated, adaptive hands, both in the case of physically-simulated experiments as well as demonstrations with a real hand.
UR - http://www.scopus.com/inward/record.url?scp=85126222817&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-95459-8_39
DO - 10.1007/978-3-030-95459-8_39
M3 - Conference contribution
AN - SCOPUS:85126222817
SN - 9783030954581
T3 - Springer Proceedings in Advanced Robotics
SP - 642
EP - 659
BT - Robotics Research - The 19th International Symposium ISRR
A2 - Asfour, Tamim
A2 - Yoshida, Eiichi
A2 - Park, Jaeheung
A2 - Christensen, Henrik
A2 - Khatib, Oussama
PB - Springer Nature
T2 - 17th International Symposium of Robotics Research, ISRR 2019
Y2 - 6 October 2019 through 10 October 2019
ER -